Temporal Proximity Query Expansion

1. A method for expanding a search query, comprising: receiving, by one or more processors, one or more initial search results from a search engine, wherein the one or more initial search results are based, at least in part, on a search query; generating, by one or more processors, a collection of events for the one or more initial search results, wherein events included in the collection of events are within a temporal proximity to at least one of the one or more initial search results; determining, by one or more processors, content of the collection of events based, at least in part, on natural language processing of at least one of the one or more initial search results included in the collection of events; determining, by one or more processors, a search suggestion based, at least in part, on the content of the collection of events; and sending, by one or more processors, the search suggestion to the search engine, wherein the search engine provides the search suggestion and the one or more initial search results to a user.

2. The method of claim 1 , further comprising: organizing, by one or more processors, a plurality of search suggestions based, at least in part, on a frequency of occurrence of the content of the collection of events.

3. The method of claim 1 , further comprising: determining, by one or more processors, an information gain for one or more features of the content of the collection of events.

4. The method of claim 3 , wherein the search suggestion is determined based, at least in part, on whether the information gain of the one or more features exceeds a predetermined value.

5. The method of claim 4 , further comprising: organizing, by one or more processors, a plurality of search suggestions based, at least in part, on the determined information gain of the one or more features.

6. The method of claim 1 , further comprising: determining, by one or more processors, one or more subsets of the content, wherein the one or more subsets of the content include at least two features selected from one or both of one or more features of the collection of events and one or more features of the one or more initial search results.

7. The method of claim 6 , wherein the search suggestion is determined based, at least in part, on whether a number of co-occurrences of the at least two features of the one or more subsets exceeds a predetermined value.